Abstract
The aim of this study was to evaluate the variation in measured glenoid inclination measurements between each of the most used methods for measuring the scapular transverse axis with computed tomography (CT) scans, and to investigate the underlying causes that explain the differences. METHODS: The glenoid center, trigonum and supraspinatus fossa were identified manually by four expert shoulder surgeons on 82 scapulae CT-scans. The transverse axis was generated either from the identified landmarks (Glenoid-Trigonum line (GT-line), Best-Fit Line Fossa (BFLF)) or by an automatic software (Y-axis). An assessment of the interobserver reliability was performed. We compared the measured glenoid inclination when modifying the transverse axis to assess its impact. RESULTS: Glenoid inclination remained stable between 6.3 and 8.5°. The variations occurred significantly when changing the method that determined the transverse axis with a mean biase from -1.7 (BFLF vs. Y-axis) to 0.6 (BFLF vs. GT-line). The Y-axis method showed higher stability to the inclination variation (p = 0.030). 9% of cases presented more than 5° of discrepancies between the methods. The manual methods presented a lower ICC (BFLF = 0.96, GT-line = 0.87) with the widest dispersion. CONCLUSION: Methods that determine the scapular transverse axis could have a critical impact on the measurement of the glenoid inclination. Despite an overall good concordance, around 10% of cases may provide high discrepancies (≥5°) between the methods with a possible impact on surgeon clinical choice. Trigonum should be used with caution as its anatomy is highly variable and more than two single points provide a better interrater concordance. The Y-axis is the most stable referential for the glenoid inclination.